This weekly update covers major developments in multi-agent AI systems, the technology that lets multiple artificial intelligence programs work together like a team.

One of the biggest stories is that enterprise companies spent $37 billion on generative AI in 2025, with a huge 3.2x increase from the year before. More than half of this money went straight into building new AI applications that companies can use right now. This shows that businesses are serious about using AI agents to get work done faster and better.

The industry made major progress on making different AI agents work together smoothly. The Model Context Protocol (MCP) became the standard way for agents to talk to each other and use different tools. In December 2025, three big companies—OpenAI, Anthropic, and Block—donated MCP to the Agentic AI Foundation, which is supported by the Linux Foundation. This means all AI agents can now understand each other better, like how everyone can understand the same language.

Researchers are also very focused on making sure multi-agent systems are safe and reliable. Scientists are worried that if one agent breaks or gets hacked, it could affect all the other agents working with it. Researchers like Sarra Alqahtani are developing tests and standards to make sure teams of AI agents can still work properly even when something goes wrong.

Companies are now building specialized AI agents that work together like a human team, instead of trying to create one super-powerful agent that does everything. Oracle announced new AI agents for their business software, and Google introduced more advanced AI agents in their products. This trend shows that the future of AI is about agents working together, not just one agent working alone.